Networking and Machine Learning techniques for Underwater Robotic and Sensor Networks (suitable for multiple students)

I am willing to discuss and supervise projects in the field of underwater sensor and robot networks.

The Oceans Systems Laboratory (OSL) [1] is involved in UK projects developing the next generation of underwater wireless sensor and robot networks [2,3]. Combining sensing, wireless networking and robots is key to tackling the challenges posed by the subsea environment. Potential applications include oceanographic exploration, monitoring of pollution and bio-diversity, monitoring and maintenance of offshore assets, disaster prevention, and assisted navigation. Communications constraints are among the most limiting factors in developing these applications, as EM waves do not propagate in water, and many standard technologies and paradigms are therefore not applicable. Coordination between sensors and multiple robots is also difficult but is a crucial requirement for allowing efficient use of available resources.

Projects of interest include:

1. Design, implementation (micropython) and evaluation of distributed sensing and data aggregation protocols to facilitate harvesting data from underwater sensor nodes using roaming AUVs through acoustic communication. This specific project is a collaboration with the National Oceanography Centre (lead), and Newcastle and Oxford universities.

2. Applications of machine learning for equalization of visible light underwater communication links [4], which can be used to transmit high bandwidth (e.g. video) data between sensor nodes and AUVs. This specific project is a collaboration with the Optical Communication and Photonic Networks Group at Heriot-Watt University.

Each topic to be investigated will cover literature survey, design/implementation and experimental demonstration. Students working on these projects will build on existing hardware prototypes and state of the art simulators/emulators, respectively, for (i) underwater acoustic sensor networks and/or optical communication, and (ii) underwater unmanned autonomous vehicles (UAVs), to simulate and visualise typical application scenarios before implementing and testing proof-of-concept systems. These projects require software programming skills appropriate for both simulation work (matlab and/or high-level languages, such as C++ / Java) and algorithmic development, especially for embedded platforms (e.g. Micropython).

[4] Chi, Nan, et al. "Challenges and prospects of machine learning in visible light communication." Journal of Communications and Information Networks 5.3 (2020): 302-309.

Supervisor name: 
Mauro Dragone
Supervisor and Deputy email addresses:
Deputy name: 
Prof. Yvan Petillot